Blanket bans, babies and bathwater: We can tell good tech from bad tech

New York City (for grades K–8) and Los Angeles (for K–12) are banning AI in the classroom, in response to the "techlash" sweeping the country, writes Fordham's Michael Petrilli. Students are using AI to "bypass" learning, concludes a new Fordham report. Fifty-seven percent of students surveyed said "high schoolers who have done no real work are getting good grades thanks to AI," and similar numbers say "most high school students are using AI in class when they are not supposed to be."

Amira's AI-powered reading took has come under attack due to privacy concerns.

Should schools ban AI? In Petrilli's debate on AI bans, many warn about throwing the baby out with the bathwater.

Michael Horn, a "techno-optimist," says AI-powered tech is far more capable than pre-AI tech. Achievement First founder Dacia Toll "is using AI to build Coursemojo, a curriculum-aligned teaching assistant for English/language arts," Horn tells Rick Hess. "She’s doing so with clear instructional design, guardrails, clarity around the importance of knowledge, and human teachers setting the rules around what good instruction, good questions, and actual student misconceptions look like."

Don't let pop culture drive policy, writes Ben Wallerstein at Whiteboard Advisors: "AI" in education means many very different things, he notes. "A literacy screener that listens to a first grader read and flags dyslexia risk earlier than any human can is not the same as Alpha School's use of AI-enabled instruction," he writes. "The work CodeAI (formerly Code.org) is doing to teach students to understand, question, and direct this technology is not remotely the same as handing a 14-year-old unsupervised access to a generalized chatbot."

Students learn less when they let a chatbot do the work, writes Natalie Wexler. But letting teachers use effective AI tools makes sense.

For example, a teacher could ask “Judy,” an AI-powered tool developed by The Writing Revolution, to "create a sentence expansion activity for fifth-graders embedded in content about the human digestive system," linking writing practice to their science lesson, writes Wexler. It's a time-saver for teachers.

Some teachers are using specialized AI tools to create “decodable” texts for beginning readers, she writes. That gives early readers more opportunities to practice on connected text.

Designing AI tutors is trickier, because students are ingenious at getting the AI to do the work for them, she writes. But she thinks "bans of tech or AI that fail to distinguish between helpful and harmful uses can end up holding students back.^"

Wexler defends Amira's tool, which listens to children reading aloud to diagnose reading problems. There's also a "tutor" version which provides personalized practice. It's "grounded in decades of research that began at Carnegie Mellon," and backed by numerous studies showing its effectiveness, she writes. It saves an enormous amount of teachers' time. But there's now public outcry because it records student's voices. Parents see that as a privacy concern.

Slowing down AI use will let schools get it right, argues Rebecca Winthrop. Even well-designed AI tools have risks, her research shows.

There are "promising applications of narrow AI, ranging from improved assistive technologies for students with learning differences to real-time translation tools that support multilingual learners," she writes. Many "involve adults using AI to better support students rather than students independently using AI systems themselves."

"Technology companies have lost parents' trust," she concludes. Educators need to include parents in the conversation about what's appropriate technology for children to use.